Abstracts Computer Science

Translation of graphics to haptic boundary representation

by Qingyuan Wang

Recently haptics became a promising research area applicable to a variety of domains, however, little attention has been paid to haptic interaction with images normally displayed graphically. Furthermore, to our knowledge, no attention has been given to how the addition of touch and force-feedback information could contribute to the task of tracing the geometrical structure of objects in images. This thesis describes the development and implementation of a Multi-Modal Image Display System (MMID system) for tracing 2D boundaries in graphic images, using a two-dimensional force-reflecting haptic interface. A haptic image model is defined to convert "what is seen" to "what is felt". Movement as an indispensable factor is an intrinsic part of the haptic image displaying process. A new performance-based rendering method for displaying image boundaries is developed to facilitate the boundary exploration, which is based on the purpose of providing users a movement guidance through an "active" haptic sensation rather than using an "impedance" technique which by definition impedes movement. An experiment is designed to evaluate tracing performance. The haptic force feedback produced by the technique introduced here, contributes to better globale performance (accuracy-speed tradeoff) and less workload for human subjects with very small levels of force intensity.